Model selection and hyperparameter tuning pose a central challenge in off-policy evaluation (OPE) because errors of fitted Q-functions and occupancy-ratio es…
机构:Stanford
来源:arXiv 2610.35748 | AI4Papers 论文推荐平台